Provider loses appeal in $4 million overpayment case involving statistical sampling
An Oklahoma-based home healthcare provider will have to return millions in overpayments to Medicare after a judge upheld the use of the controversial statistical sampling method in the case against it.
San Bois Health Services Inc. was asked in 2009 by zone program integrity contractor Health Integrity to produce records supporting just 56 home health claims filed between 2007 and 2009. Health Integrity determined that 46 of the 56 sampled claims did not meet Medicare's home health coverage criteria. From that small sample, the contractor extrapolated the result to claim that the provider was overpaid roughly $5 million.
San Bois filed an appeal of the overpayment determination, arguing that the statistical sampling method was “irreparably flawed” since it relied on the same “seed number” to generate the number of claims to be selected, according to court records.
Statistical sampling has come under fire under recent years for its “novel” method of prosecuting cases, with the American Health Care Association slamming it as an improper “sledgehammer.” False claims against another long-term care operator in a case that gained attention for its use of statistical sampling were dismissed in August.
In the San Bois case, Health Integrity eventually recalculated the overpayment to $4.1 million. In 2014, a Medicare appeals council reversed an earlier administrative law judge's decision, determining that the use of the same seed number was “simply not a ‘flaw' in the sampling process” under Centers for Medicare & Medicaid Services guidelines.
In a ruling posted Monday, the U.S. District Court for the Eastern District of Oklahoma found that Health Integrity had met the federal requirements for proper statistical sampling.
“There is substantial evidence — undoubtedly more than a scintilla — in the administrative record that the ZPIC satisfied the MPIM requirements for a probability sample and a stratified random sample,” wrote Judge Ronald A. White.
White affirmed the government's decision in the case, agreeing with experts that a sample isn't invalid simply because a different or more precise method is available. San Bois' motion to reverse the finding was denied, and the case was dismissed.